Determinants of pre-interaction swift trust and long-term adoption of care robots: the role of mind attribution and perceived similarity in eldercare robotics
Mahed Maddah, Shirin Hasavari, Shadi Esnaashari, Pouyan EsmaeilzadehPurpose
The global eldercare crisis, driven by population aging and caregiver shortages, has accelerated interest in care companion robots. However, their integration into care settings depends partly on whether older adults are willing to trust unfamiliar robotic agents. Existing research has examined robot design, functionality and cultural acceptance, but has not fully integrated mind attribution and perceived relational alignment in explaining prospective trust and adoption intentions. This study therefore applies relational demography theory (RDT) as its primary framework, complemented by the computers are social actors (CASA) paradigm and self-congruence theory, to examine how perceived mind attribution, caregiving fit, and pre-existing attitudes are associated with anticipated trust in a care robot.
Design/methodology/approach
Using a cross-sectional, video-based survey, the study collected data from 260 USA adults aged 55 and older who evaluated Pepper in a standardized caregiving scenario. Participants viewed informational demonstrations that approximated an initial, observational encounter and then completed measures of seven constructs. The hypothesized model was tested using covariance-based structural equation modeling in AMOS 26.0, with three item parcels per construct (21 composite indicators).
Findings
Perceived mind attribution (β∧ = 0.39), caregiving fit (β∧ = 0.37) and general attitudes toward robots (β∧ = 0.31) were positively associated with anticipated pre-interaction trust. That trust was associated with long-term adoption intentions through two significant indirect pathways: perceived improvement in well-being (β∧ = 0.27) and preference for robotic over human caregivers (β∧ = 0.25). The model accounted for 61% of the variance in adoption intentions.
Originality/value
This study provides preliminary empirical support for applying selected RDT similarity mechanisms to prospective human-robot evaluation in eldercare and integrates mind attribution and pre-interaction trust into that framework. Because participants evaluated one robot through video-based exposure without direct interaction, the trust construct captures anticipated, pre-interaction trust rather than interaction-based swift trust. Accordingly, the contribution is bounded to prospective evaluations of a specific robot stimulus and should not be interpreted as validating RDT or swift-trust mechanisms across interactive human-robot settings. Within this boundary, the model clarifies how mind attribution, caregiving fit and pre-existing attitudes are associated with adoption intentions through well-being and comparative-preference pathways.